Patterns in Lottery Numbers

markmcb writes "Most everyone is familiar with the concept of the lottery, i.e., random numbers are selected and people guess what they will be for a cash prize. But how random are the numbers? Matt Vea has conducted a pattern analysis of the MegaMillions lottery, which recently offered a sum of $370M (USD) to the winner. Matt shows that the lottery isn't as random as it may seem and that there are 'better' choices than others to be made when selecting numbers. From the article, 'A single dollar in MegaMillions purchases a 1 in 175,711,536 chance of landing the jackpot ... a player stands a mildly better chance of winning a partial prize through the selection of weighted numbers.'" Includes some excellent charts of his analysis.

Comments

Re:And yet, one truth escapes the analysis

[Red Flayer]

You're already losing by buying the ticket.

From an arithmetical perspective, yes. From a more subjective standpoint, no.

Instead of using absolute dollar figures for your analysis, you should use lifestyle impact.

e.g. One dollar a week == no lifestyle impact; $370MM payout == off the charts lifestyle impact.

This is why people will continue to play the lottery, even if mathematically it's a poor choice.

Lottery vs. poker

[darjen]

A single dollar in MegaMillions purchases a 1 in 175,711,536 chance of landing the jackpot ... a player stands a mildly better chance of winning a partial prize through the selection of weighted numbers.

As an amateur poker player, I'd say I have a better chance of winning the world series of poker. There is absolutely no reason for the lottery to be the only form of legal gambling, while poker remains illegal. Other than the states want to maintain their monopoly on gambling. God forbid I should spend $30 on a tournament buy-in, while someone down the street buys 30 legal lottery tickets. Especially since I have a pretty good shot at making more than my $30 back.

Fl. lottery...

[i.r.id10t]

Fl. Lottery used to print which machine would be used for a particular drawing, and old drawing numbers - with machine that was used - in a nice column format (probably real easy to rip out of a web page, etc).

Now I'm not a math guy, but I *know* that all 50whatever balls aren't identical. Should there be some (very) slight preference for a particular ball, or set of balls, over a series of draws over time? I seem to remember an article in a Dragon magazine about dice and using math to find out if yours had a preference for a particular number...

Re:How about ANY random numbers?

[AmigaHeretic]

I wonder how much ANY person would see ghosts if one tried hard enough to find them at the Waverly Hills Sanatorium?

Re:You can't lose if you don't play

[zerocool^]

One of the slogans for the Illinois lottery used to be "you can't win if you don't play", but I figured every time I didn't play, I won $1. Its both stupid and ironic that many of the same people bitching about taxes pay this voluntary tax!

At the astronomical odds against winning, I figure my chances of finding a winning ticket on the ground are only marginally worse than my chances of buying a winning ticket. So rather than give extra money to the government so it will be funnelled to politically connected rich people, I just watch the ground.


I play the Mega Millions every time the numbers are drawn. We have a standing office pool with 4 of us, and we play 4 numbers every time the balls are dropped, the same 4 sets of numbers each time.

What have we won? Well, one time we won $150. But aside from that, we know that the lotto money goes into the public school system. So, in effect, I'm donating $8 - $9 minus administrative fees to the school system every month.

I'm ok with that.

And on the off chance that I ever win the lotto, none of you will ever hear from my fat ass again.

~Wx

Re:Missing data

[Red Flayer]

A lot of lottos that offer a computer 'quick-pick' are set up so that the quick-pick only chooses numbers not already taken. Of course, someone could namually choose your set of numbers after you buy your ticket, but if you want to maximize your potential earnings, always use the quick-pick (for lottos that obey this rule).

College stats course

[rueger]

Of all of the courses I took over the years, my college Statistics course stands out. The teacher was the kind of guy who would spend an evening at his kitchen table flipping pennies just to prove to himself that yes, ultimately they would trend towards a 50/50 distribution. For fun.

What stuck with me though we a couple of ideas:

• What happened in the past has no influence over what will happen in the future. You may have flipped seventy-five heads, but the odds of the next penny landing head or tails is still 50/50.
• The decision whether to make a bet in any game is based on a variety of factors - the size of the bet, the size of the possible prize, and the odds of winning. No one of those three is enough to make a decision, you need to know all three.

In life, as in games, you have to make decisions based on real odds, not of conjecture or media speculation. Where's the bigger risk? A falling meteor, or getting hit by a city bus? Food poisoning or a terrorist attack?

Re:I've known people that used patterns

[greg1104]

Overall, though, the guys that I have known who have done this sort of stuff have won, though. Not tons, maybe an average of a couple hundred dollars per month at best with the risk of losing thousands

It sounds like a classic case where someone is staying just ahead of Gambler's Ruin--so far. If you don't properly model what happens when you run out of money, there are any number of gambling systems that seem like they work in the short-term, but in reality don't if you run them long enough. The simplest one is the Martingale.

An interesting lotto winner story

[Anonymous Coward]

The famous linebacker for the Dallas Cowboys in the 1970's, Thomas "Hollywood" Henderson.

He blew every cent he earned playing football on drugs and was arrested on sex and drug charges. When he was broke he found religion and got cleaned up. Later, he hit the lottery for $28 million.

My first question

[artifex2004]

Does this guy know there are multiple ball sets in these big games, that are "randomly" selected before each drawing?

Looking at statistical weighting of ball frequency in a huge aggregate is fine. But I'm guessing (but not betting) you can't translate it directly into predictions in the short term, because of this random set issue. Now, if you want to make multiple analyses based on the specific historical performance of each set, then you have something more practical. But then you need to buy multiple sets of tickets based on the multiple predictions you get. I don't think his data source will tell him what ball set was used.

(Early on, when Lotto Texas was new, I used to guess 2 or 3 of the numbers based on simple frequency analysis, in every single drawing. But they kept adding ball sets. And Texas did tell people which ball set they used. After the fact, of course. And then I think they changed the number of balls picked. So I gave up my paper trials.)

If your office mates win, and you're not in it...

[mckwant]

I'm just saying. I don't care about $5/month, but if those guys win, and I'm still at this $@#%ing desk, that'll be bad.

It's not gambling, it's "gloating ex-co-worker insurance."

Lottery and the Mob

[brit74]

Back in the early 20th century, the mob used to run a kind of informal lottery. It was called a "numbers game". There were places (like barbershops) where people could pick a number from 1 to 1000, and if their number came up, they won. The mob typically paid out between 800-to-1 and 600-to-1. This meant that the mob paid out 60-80% (and kept 20-40%) of the money people initially paid. On the other hand, most state lotteries only pay out about 50% of earnings, making them a worse bet than going with the mob.
http://en.wikipedia.org/wiki/Numbers_game
http://marketplace.publicradio.org/features/underground/1114undergroundpm.html

Re:You can't lose if you don't play

[Phat_Tony]

Mod parent up. This is a common misunderstanding, or perhaps I should call it a government lie. It's Enron accounting.

They try to encourage people to play the lottery by telling them the money all goes to schools. So they direct the whole of lottery income to the schools, on paper. But the school budget is planned in budget meetings, and the ammount the lottery brings in has no effect on school budgets. For every extra dollar brought in by the lottery, that's one less dollar of general fund money that goes to the schools, and one more dollar of total money in the government's general fund.

They can write out the accounting anyway they like to and say, "see, these dollars went here," but dollars are fungible. At the end of the day, the acid test to see if it means anything to say "lottery money goes to schools" is to see what the marginal effect on school funding is of purchasing lottery tickets. That effect is 0.

Re:And yet, one truth escapes the analysis

[Ced_Ex]

Do you know what that dollar used for a lottery ticket buys you?

Basically it buys you a dream, and an opportunity to share your dreams with your friends/co-workers who are in on the ticket. Really no different from buying a piece of art for a conversation piece.

At work, we buy lottery tickets together when the jackpot is huge, and then we get to talk about the things we'd do if we won, which in effect buys me a mental vacation from the reality of having to work for a living. For a dollar every now and then, it's worth it. Besides, a losing ticket just buys social assistance projects where I'm from anyways, so it's not really a big loss.

Re:And yet, one truth escapes the analysis

[Scrameustache]

Also, for 1 dollar a week you are buying a mild emotional high when checking the numbers and a mild emotional low when you find out that math pwned you again.

Yes, but for less than one dollar a day (that's less than the cost of a cup of coffee), you can help change a child's life. Imagine the feeling of joy you'll get when you receive a personalized card in the mail from the child you sponsor. Call the number on your screen right now to make a donation. My dad's old boss went to visit his sponsored child for his vacation once.
Turned out the kid had been dead for years. Someone kept collecting the money, though, and wrote letters.

Interesting scam based on that

[A nonymous Coward]

I was quite amused when I first heard of it. I d onot know if it has ever actually been done, however.

Pick some random stock, send email to 2*10 suckers that it will go up, and to an equal number of suckers that it will go down.

Whichever direction it moves, divide that batch of suckers in two, pick some other random stock, send half email taht it will go up and half email that it goes down.

Repeat until you have only a few suckers left. They will see you as a genius who has correctly predicted the last {n} stock moves correctly. It may be the final sucker, or the last 2, or 4, etc.

Now tout a stock you have just bought and make some money!

Re:How about ANY random numbers?

[myowntrueself]

Its not too hard.

One method is to take each pair in the sequence as coordinates in a plane and plot them.

Eg, you have a dripping tap. You take the time between drips as your number. Then two successive intervals get plotted on a plane and you can start finding a resulting pattern.

I once did this with pseudo-random number generators as provided with various libraries.

The results were interesting.

Each different PRNG would produce a distinct pattern and the same pattern every single time regardless of whether you initiated with a different seed value. Its like a fingerprint; you look at the resulting pattern and you can say which PRNG produced it.

This was in the '90s and the Borland 'turbo pascal' PRNG produced a pattern that looked a LOT like a space-shuttle :)